parse a string (Cadence Allegro Netlist) to dictionary

Rhodri James rhodri at wildebst.demon.co.uk
Thu Nov 5 19:42:39 EST 2009


On Thu, 05 Nov 2009 20:02:53 -0000, Leland <lelandpeng at gmail.com> wrote:

> Hi,
>
> I always use readline(), strip(), split() and so on to parse a string.
> Is there some elegant way to parse the following string into a
> dictionary {'50MHZ_CLK_SRC' : 'U122.2, R1395.1'}?
>
> NET_NAME
> '50MHZ_CLK_SRC'
>  '@TEST_LIB.TEST(SCH_1):50MHZ_CLK_SRC':
>  C_SIGNAL='@test_lib.test(sch_1):\50mhz_clk_src\';
> NODE_NAME     U122 2
>  '@TEST_LIB.TEST(SCH_1):PAGE92_I223 at INF_LOGIC.CY2305(CHIPS)':
>  'CLK2': CDS_PINID='CLK2';
> NODE_NAME     R1395 1
>  '@TEST_LIB.TEST(SCH_1):PAGE92_I232 at INF_RESISTORS.RESISTOR(CHIPS)':
>  'A': CDS_PINID='A';

Here's an inelegant way:

**** CODE ****

results = {}
net_name_next = False
net_name = None
node_names = []
for line in sourcefile:
   line = line.strip()
   if line.startswith('NET_NAME'):
     if net_name is not None:
       results[net_name] = ", ".join(node_names)
     net_name = None
     node_names = []
     net_name_next = True
   elif net_name_next:
     net_name = line.strip("'")
     net_name_next = False
   elif line.startswith('NODE_NAME'):
     node_names.append("{1}.{2}".format(*line.split()))

# Last time through
if net_name is not None:
   results[net_name] = ", ".join(node_names)

**** END CODE ****

If you're prepared to allow the dictionary values to be lists rather than  
strings, we can squash that down:

**** CODE ****

results = {None: []}
net_name = None
for line in sourcefile:
   line = line.strip()
   if line.startswith('NET_NAME'):
     net_name = sourcefile.next().strip(" \t\n'")
     results[net_name] = []
   elif line.startswith('NODE_NAME'):
     results[net_name].append("{1}.{2}".format(*line.split()))
del results[None]

**** END CODE ****

If you can guarantee that you won't meet a NODE_NAME before you see a  
NET_NAME then you can lose the messing about with results[None].  Having  
spent today picking up the pieces from an assumption that's rotted after  
several years, I'm not feeling that brave.  A slightly less messy version  
of that would be:


**** CODE ****

results = {}
entry = []
for line in sourcefile:
   line = line.strip()
   if line.startswith('NET_NAME'):
     entry = []
     results[sourcefile.next().strip(" \t\n'")] = entry
   elif line.startswith('NODE_NAME'):
     entry.append("{1}.{2}".format(*line.split()))

**** END CODE ****

I'm a little dubious about doing mutable magic like this without copious  
comments, though.

-- 
Rhodri James *-* Wildebeest Herder to the Masses



More information about the Python-list mailing list